Many a times we come across situations in which during moderate traffic hours, traffic light at a junction follows a constant fixed cycle of time T thereby reducing the easy and continuous flow of traffic across a signaled junction. We in this paper focus on overcoming this inherent defect of non-intelligent traffic control system by incorporating intelligent ‘traffic level sensing algorithm’ based on inputs from RFID sensors placed at certain fixed places around the junction. Though much research work had been done in this field using game theory approach, ”Ref.[1]”associative mapping technique, graph model, etc., none of them prescribes an accurate way of determining exact traffic levels at any junctions. Though these approaches provide a very efficient way of managing traffic using mathematic models, most of them are based on repetitive learning, where the control system takes a long time in learning the traffic patterns spread across days, months and years. Our work basically involves calculating the traffic queue on any road in a junction thereby eliminating the repetitive learning process. We have also developed a simple algorithm that takes decisions based on the input data from the RFID sensors.